MIT Unravels the Secrets Behind Collective Intelligence – Hint: IQ Not So Important
When it comes to a successful group, the easiest way to ensure victory may be placing women on the team. MIT's Center for Collective Intelligence seeks to understand how humans get better (or worse) at solving problems as they work together. They studied hundreds of people working in small groups and found that they could determine a "C factor", a key statistic that would predict if a group could perform well in a variety of tasks. C factor was more important in determining group success than the individual IQs of the people in the group. In other words, having a successful team isn't just about having smart people, it's about having people who will work together well. And what gives a group a high C factor? Women. Well, to be more precise, a high level of social sensitivity and willingness to let everyone talk equally. As forms of collective intelligence grow in importance, as we see with crowd-sourcing projects like Wikipedia, social search engines, and the scientific community, the value of socially aware individuals is going to arise as well. Is the future going to be inherited by the peacemakers?
MIT's research into measuring collective intelligence was lead by their own Thomas Malone in partnership with Carnegie Mellon's Anita Woolley. The results were published in late 2010 in Science Magazine and have since been discussed in over 30 major press agencies. Everyone loves to talk about how women make teams better, probably because it fulfills some of our most beloved (reviled?) 20th Century Western stereotypes. "Women Are the Key To a Successful Team" makes such a great headline, doesn't it? Despite my own use of that trope (mea culpa), Malone and Woolley didn't find that women per se were the key to a good C factor. It's just that social sensitivity, which was overwhelmingly the leading ingredient in high C factor, was overly correlated to women. In fact, when they controlled for the number of women in a group, it was shown that it was the emotional sensitivity scores which won out. So adding a woman to a team no more guarantees higher success rates than adding an Aaron to your dance party will make it a success. It's the higher levels of those particular skills amongst that group which make the difference, not the presence of that group itself.
Woolley explains this, and many other parts of their collective intelligence research, in the following video:
When Malone, Woolley and the rest of the MIT team set out to measure collective intelligence, it wasn't even clear if such a thing could statistically be shown to exist. Sure, people have spent years measuring how groups perform a task, but no one was really asking if a group that was good at one task would somehow be good at most tasks. That's something we tend to associate with individual IQ, which is often used as a measure for one's universal ability to handle complex problems. Does group IQ, or collective intelligence, really exist? If so, is it just a result of the average individual IQs in the group?
It turns out that a group is more than the sum of its IQs. In two studies, the MIT team watched how small groups (2-5 people) handled complex problems. They made these groups tasks like visual puzzles, brainstorming,making collective moral judgments, and negotiating over limited resources. There were also two criterion projects (study 1 had the team play checkers against a computer, study 2 had the team work on architectural design) where individual IQs were a solid predictor for success. After these studies (which had teams working together for up to five hours!) Malone and Woolley found that some groups were indeed better at almost everything they tried. While maximum and average individual IQ could weakly predict a good group, they found there was a much more important collective intelligence (C) factor that determined a group's success better. If a team had a high C factor they tended to do well in many different activities, even when compared to groups with higher average IQ.
Along with discovering the C factor, Malone and Woolley also loosely determined what went in to creating it. Surprisingly, many common sense variables like group cohesion, motivation, and satisfaction simply didn't seem to matter much. Instead, MIT determined that groups with a high C factor were mostly defined by:
- the high average social sensitivity of group members
- a high rate of sharing who gets to communicate
- more females
As I've said before, the number of females could be controlled for and the other two ingredients would still shine through. Shared communication is pretty easy to understand as a concept, and it may make sense that it leads to better collective intelligence. A single smart person dominating a group isn't allowing the team to benefit from their shared expertise and insights as well as a a group that asks for input more evenly. Getting everyone in the team to participate, or at least allowing them the opportunity to share, is key to harnessing collective intelligence. MIT studied not simply talking, but nonverbal communication as well, so it may be that this openness corresponds to each person feeling listened to as well.
Social sensitivity, as measured by these experiments, is really about understanding what people around you may be feeling based on small cues. The classic test used to measure social sensitivity is the "Reading the Mind In the Eyes Test" which you can take online. Again, it makes sense that this skill was evident in furthering collective intelligence. Being able to quickly look around a table and determine that one person is very confident in their knowledge, another is impatient with the task, and another wants to share information but isn't feeling able is going to be a big advantage in getting the group working together at maximum potential.
While it's still very early in the history of measuring collective intelligence to understand exactly how Malone and Woolley's research will impact the world, there are clear applications in almost every business environment. Eventually, corporations may be able to use quick social sensitivity tests to better determine how best to form teams out of a pool of employees. There's also the possibility that collective IQ can be improved through training much more than individual IQ. If so, we may be able to take any rag tag group of individuals and teach them how to be a badass problem solving strike team.
Perhaps most importantly, now that we have a fledgling measure of collective intelligence, we may be able to understand which of the emerging forms of crowd-sourcing are really worthwhile investments. Does it make more sense to have a million people working on a problem, if their cooperation is terrible? Maybe a project with just 1000 well coordinated participants will be better? The measurement of a C factor changes the answers to those questions from guesswork into quantitative analysis. With enough research the crowds of tomorrow may be optimized for the best possible amounts of collective intelligence. Not just huge amounts of thought-power, but efficiently organized huge amounts of thought-power. Talk about working smarter not harder.
Latest posts by Aaron Saenz (see all)
- How Today’s Jungle of Artificial Intelligence Will Spawn Sentience - August 9, 2016
- Welcoming Your New Robot Overlords - July 16, 2013
- Steve Jurvetson Considers Emerging Fields, Singularity University - September 30, 2012